Low-Complexity Band-Toeplitz Linear Precoding inMassive MIMO system
首发时间:2015-12-02
Abstract:Massive multiple-input multiple-output (MIMO)system is getting more and more attention lately because ofit’s substantial improvement in spectral efficiency. However, thechallenge for Massive MIMO is the increased complexity of signalprocessing in the case of large antenna arrays. For example, thenear-optimal regularized zero-forcing (RZF) precoding is difficultto achieve when the number of antennas is huge, because ofthe inversion of large dimensional matrices. We develop a newimproved algorithm for RZF precoding in this paper. At thebeginning of this algorithm, the ordinary conjugate symmetricmatrix inversion is replaced with a Toeplitz matrix inversion. TheToeplitz property is based on the approximate orthogonality andsymmetry between massive MIMO antennas. Then the Toeplitzmatrix is further evolved into a banded Toeplitz matrix, as mostof the Teoplitz matrix elements tend to zero. Finally the bandedToeplitz matrix inversion is solved efficiently by the improvedLevinson-Durbin algorithm we proposed. Further by simulationat the end of this paper, we show that the per-user rate loss ofour algorithm compared to that of RZF is relatively small andaccepted with the large complexity reduction.
keywords: massive MIMO precoding inversion Toeplitz band
点击查看论文中文信息
Massive MIMO系统下基于带状Toeplitz矩阵的低复杂度线性预编码
摘要:大规模MIMO系统因其对频谱效率的巨大提升,正逐渐受到越来越多的关注。然而,大规模MIMO面临的一个挑战是随着天线数目的增长而带来的信号处理的复杂度。例如,因为求逆矩阵的维度非常大,传统的接近最优的 regularized zero-forcing (RZF) 预编码变得很难实现。本文提出了一种基于 RZF 预编码的优化算法。本文提出的算法首先用 Toeplitz 矩阵的求逆替换 RZF预编码中的求逆矩阵,再进一步根据对称性优化此 Toeplitz 矩阵为带状矩阵,然后用改进的 Levinson-Durbin 算法高效地解决矩阵的求逆。本文的最后仿真验证了此优化算法大大降低了计算复杂度,相应的性能损失却很小。
基金:
论文图表:
引用
No.4665770111736114****
同行评议
勘误表
Massive MIMO系统下基于带状Toeplitz矩阵的低复杂度线性预编码
评论
全部评论0/1000